摘要

Practical plans for agricultural water use within multiobjective frameworks require feasible solutions that meet the objectives of competing interests and this is the root of common decision problems that plague water resources systems. To find the best solutions among the set of feasible solutions, decision-makers can analyze their worth functions, which is the goal of the surrogate worth trade-off (SWT) method, which enables assessment of the worth functions after solutions that are optimal to separate groups are analyzed. The SWT method uses Lagrange multipliers to determine the set of Pareto optimal solutions and requires the exact equation of each objective function and its derivative or gradient. This is normally impractical in watershed scale problems because each objective function comprises a set of interactive physical and hydrological equations, but the problem can be partially overcome by incorporating a genetic algorithm. This approach was applied to a case study of California%26apos;s San Joaquin River watershed to approximate optimum rates of reduction in agricultural water allocations for environmental purposes. In the case study, decision-makers were aided in assessing their worth functions on the basis of the optimal solutions presented to them. The genetic algorithm optimization tool was linked to a watershed simulation model using the soil and water assessment tool (SWAT) to simulate streamflow and salinity. Model results showed that SWAT provides satisfactory predictions for salinity, which can be used in the trade-off analysis. The compromised rates of agricultural water allocations resulted in a significant increase in the system%26apos;s reliability and decreased its vulnerability to salinity.

  • 出版日期2014-8